Pale Moon SSE

[Mercury]

The SSE build of Pale Moon is now available here. As of this writing, v25.7.3 as well as v26 beta 3 are available for download. Since this version is built from the Atom source branch, it supports Windows XP, but lacks Web Developer tools.

Many thanks to Moonchild for providing hosting space.

Please report any problems or issues specific to the SSE builds in this thread. Questions and comments are also welcome.

For those few of us with museum-grade hardware and a desire for a sane browser: Enjoy!

[Alt+F4]

YES!

Let me be the first to congratulate you. Brilliant job. I did a clean install (.exe file), working flawlessly.

Thanks a million, Mercury. It’s good to know that someone still cares about old PC users.

Keep up the excellent work!

P.S. Also a big thanks to Moonchild and Matt!

[Mercury]

Version 25.8.0 is now available.

Now with extra complier-flag goodness! (Added "-GFLs" to the options for "whole program optimization". The mainline build uses this option; it was omitted in the previous SSE build by mistake.)

I did a clean install (.exe file), working flawlessly.

Sweet.

And thank you for your thanks.

[New Tobin Paradigm]

Version 25.8.0 is now available.

Now with extra complier-flag goodness! (Added "-GFLs" to the options for "whole program optimization". The mainline build uses this option; it was omitted in the previous SSE build by mistake.)

I did a clean install (.exe file), working flawlessly.

Sweet.

And thank you for your thanks.

You shouldn't be using /GL for builds with NT 5.x win32 subsystem targets..

[Alt+F4]

Version 25.8.0 is now available.

I did a clean install (.exe file), working flawlessly.

Sweet.

And thank you for your thanks.

You're most welcome. And thank you once again!

Brilliant work. Keep it up!

[Mercury]

You shouldn't be using /GL for builds with NT 5.x win32 subsystem targets..

I have encountered no problems with this build. It works stably, and is just as fast, if not faster, than my previous build. My research didn't turn up any warnings against LTCG specifically for NT5 targets.

LTCG is a perfectly good optimization. Saying it causes problems implies that either the Pale Moon source or Microsoft's build tools have a serious but subtle bug. I certainly don't claim the impossibility of such a thing but I'm not going to be afraid of a boogeyman.

This build is intended for pre-SSE2 CPUs which are already underpowered; I'll take every optimization I can get.

Unless a concrete problem report surfaces — or MC forbids it, of course — I'll keep this build published and continue to apply the optimization in the future.

If you can be more precise about what the problem is, I can verify whether it occurs.

[Alt+F4]

@Mercury

Just to let you know, it's working great so far.

[Mercury]

Just to let you know, it's working great so far.

Great to hear! Please keep the feedback coming.

I did talk to MC and he described the problem as "crashes that can't be reproduced, startup/shutdown issues of the application, inexplicable network delays, graphics glitches... None with a defined cause". Of course, I haven't seen anything like this.

Hey, if it's a complier bug, maybe compiling for SSE avoids the problem?

I'm willing to rebuild with different options, of course, but I'd need at least one actual problem report: Instability with 25.8.0.SSE that did not occur with 25.7.3.SSE.

[Alt+F4]

Just to let you know, it's working great so far.

Great to hear! Please keep the feedback coming.

Will do!

I did talk to MC and he described the problem as "crashes that can't be reproduced, startup/shutdown issues of the application, inexplicable network delays, graphics glitches... None with a defined cause". Of course, I haven't seen anything like this.

Me neither. I'm keeping my fingers crossed to stay that way.

[Moonchild]

If you're going to take the opportunity to quote a private message publicly (bad form, please don't) then at least provide all the context that I used to explain this.
There is no guarantee that a current problem-free build will remain problem-free. If it runs smoothly,right now, then that's great! This is drawn from our past experience over the years, so we're not pulling this out of our hats and recommend not using /GL for these builds. But it does mean that if problems crop up, it'll be on your own heads.

It's also quite possible that offending code (e.g. time-critical telemetry routines) is simply no longer present in our code and that it can safely be used at this point, but it's a trial&error deal. Only Microsoft knows the inside of their compiler.

What I said was:

I'm fine with you taking the risk, as long as you're prepared for the possible fall-out. It's your build, after all -- so your responsibility.

I know "less stable" is vague, but so are the intermittent problems we've seen: crashes that can't be reproduced, startup/shutdown issues of the application, inexplicable network delays, graphics glitches... None with a defined cause. It seems to hit a compiler bug or several that creates timing-sensitive problems; something you may not see and your neighbor may not see, but someone one a different network, different class of PC or different combination of drivers and hardware may run into. It's also possible that your current build won't exhibit these issues, but the same build of the next point release does, since it's a compile-time bug as far as I've been able to gather. It's a strong recommendation to not use it, but the final call is yours.

[Mercury]

If you're going to take the opportunity to quote a private message publicly (bad form, please don't)

It was tech-talk, and I only quoted bug symptoms so that others could know what to look for. I didn't mean any offense; I just assumed things of that nature are okay to repeat. I won't do it again.

There is no guarantee that a current problem-free build will remain problem-free. If it runs smoothly,right now, then that's great! This is drawn from our past experience over the years, so we're not pulling this out of our hats and recommend not using /GL for these builds. But it does mean that if problems crop up, it'll be on your own heads.

That's fair. I didn't want to imply that I don't believe you or Tobin. But I saw no evidence of this issue, so I figured it may have been something that's been eliminated through code changes (like you just said), or a quirk of the specific build environment you were using, or a voodoo curse, or something else which makes it no longer applicable.

There's a very small contingent of people using this build. I take full responsibility for using them as my guinea pigs.

So, yes, to make it as clear as possible:
Version 25.8.0.SSE was built, by me, with a compiler flag that is discouraged by the PM devs in this case, and the resultant build has a higher risk of stability issues.

This is my decision, made because I think the probable speed improvement is worth it. The underlying issue may not exist in this specific build.

[Moonchild]

Off-topic:

It was tech-talk, and I only quoted bug symptoms so that others could know what to look for. I didn't mean any offense; I just assumed things of that nature are okay to repeat. I won't do it again.

It's just a rule of thumb, really. Either ask if it's OK to post a private message sent to you publicly, or make a re-written synopsis to relay to general thoughts behind the message, but don't directly quote PMs. It's just bad form to do so. I'm not upset or anything, just pointing out that in general, I really don't like this; it's considered a "faux pas".

There's a very small contingent of people using this build. I take full responsibility for using them as my guinea pigs.

Certainly fair enough -- And if you do run into issues (that actually may or may not be present) it's always easy enough to rebuild.

[Mercury]

Version 25.8.1 is now available for download.

[RJARRRPCGP]

I did talk to MC and he described the problem as "crashes that can't be reproduced, startup/shutdown issues of the application, inexplicable network delays, graphics glitches... None with a defined cause".

Sounds like bad caps on socket 462 and 478 period motherboards. If any caps oozing or bulging, that's the cause right there.

A bad overclock also causes those symptoms, random and frequent crashes.

[Alt+F4]

Version 25.8.1 is now available for download.

Thank you I really appreciate your work. Keep it up!

[Mercury]

Sounds like bad caps on socket 462 and 478 period motherboards. If any caps oozing or bulging, that's the cause right there.
A bad overclock also causes those symptoms, random and frequent crashes.

Pretty sure MC or Tobin would've noticed a hardware issue like that.
I may not see eye-to-eye with them on everything, but I believe them when they say this was a real compilation problem.

So far there have been no reports of instability. Both V25.8.0.SSE and v25.8.1.SSE were built with -GL, so I take no news as good news.

Out of curiosity, if MC reads this and has the ability, could we get a report on how many downloads of the SSE version there have been?
I seem to have one die-hard fan — wondering if there are any others.

[Alt+F4]

@Mercury

Now you have a responsibility to learn how to manage your fame and celebrity status.

[Mercury]

Version 26.0.2 is now available for download. (Version 26.0.0 has been skipped.)

[Alt+F4]

Congrats Mercury, you did it yet again!

Great job, keep it up and thank you!

Also a big thanks to the PM team!

[Reino]

Thanks a lot, Mercury, for the continued support for my old AMD Athlon XP3200+.
I just updated from v25.7.3 to v26.0.2 and so far it's working great, BUT I noticed all my tabs looked different, so I start searching and found this in about:config:

Code: Select all

extensions.colormytabs@jetpack.activeTabBrightness;         70
extensions.colormytabs@jetpack.activeTabFadingPower;        5
extensions.colormytabs@jetpack.activeTabFadingRange;        5
extensions.colormytabs@jetpack.activeTabOpacity;            100
extensions.colormytabs@jetpack.activeTabSaturation;         40
extensions.colormytabs@jetpack.allowColorBrightnessFixes;   1
extensions.colormytabs@jetpack.boldActiveTabTitle;          true
extensions.colormytabs@jetpack.hoveredTabBrightness;        65
extensions.colormytabs@jetpack.hoveredTabFadingPower;       5
extensions.colormytabs@jetpack.hoveredTabFadingRange;       5
extensions.colormytabs@jetpack.hoveredTabOpacity;           100
extensions.colormytabs@jetpack.hoveredTabSaturation;        35
extensions.colormytabs@jetpack.improveTabTitleVisibility;   true
extensions.colormytabs@jetpack.inactiveTabBrightness;       60
extensions.colormytabs@jetpack.inactiveTabFadingPower;      5
extensions.colormytabs@jetpack.inactiveTabFadingRange;      5
extensions.colormytabs@jetpack.inactiveTabOpacity;          100
extensions.colormytabs@jetpack.inactiveTabSaturation;       30
extensions.colormytabs@jetpack.showActiveTabIndicationBar;  true
extensions.colormytabs@jetpack.tabDefaultColor;             #FBFCFD
extensions.colormytabs@jetpack.tabFadingColor;              #F0F0F0
extensions.colormytabs@jetpack.tabFadingStyle;              3

I haven't removed my add-ons before uninstalling v25.7.3, but Color My Tabs wasn't one among the ones I had installed. Now with v26.0.2 I also can't seem to uninstall the add-on, because it's not listed anywhere. Have you perhaps hard-coded this add-on in the installation setup?